Exposure of neonates to ochratoxin A: first biomonitoring results in human milk (colostrum) from Chile
- First Online:
- Cite this article as:
- Muñoz, K., Campos, V., Blaszkewicz, M. et al. Mycotox Res (2010) 26: 59. doi:10.1007/s12550-009-0040-0
- 124 Views
The mycotoxin ochratoxin A (OTA) and its metabolite ochratoxin alpha (OTα) were determined in milk and blood from nine lactating women who provided samples soon after delivery at a hospital in southern Chile. The analytical method applied liquid–liquid extraction with chloroform, and in the case of blood, an extra purification with solid phase extraction prior to HPLC analysis with fluorescence detection. OTA was detected in all human milk samples, with an average concentration of 106 ± 45 ng/L (range 44–184 ng/L). Levels of OTα were 40 ± 30 ng/L (LOQ 40 ng/L), but increased considerably upon enzymatic hydrolysis with ß-glucuronidase/sulfatase (up to 840 ± 256 ng/L) in human milk. By contrast, there was no evidence for conjugates of OTA. The data on OTA in breast milk and levels reported in blood from women in Chile are indicative of an efficient lactational transfer of the mycotoxin. Infant exposure to OTA was estimated by considering their daily OTA intake with human milk at early stages of nursing. For the majority of milk samples, the calculated OTA intake of infants exceeded the tolerable daily intake (TDI) of 5 ng/kg body weight (bw)/day proposed by the Nordic Expert Group, and infant exposure approached the provisional tolerable doses of 14–16 ng/kg bw/day suggested by the Joint FAO/WHO Expert Committee on Food Additives (JEFCA) and by EFSA for adults. The present study documents and confirms the presence of OTA in human milk at levels where the TDI can be exceeded. These results point out the need to continue food and biological monitoring and to develop strategies, e.g. dietary recommendations to pregnant and lactating women, aimed to reduce OTA exposure in early periods of life.
KeywordsBiomonitoringHuman milkHPLC analysisOchratoxin ATolerable daily intake
The mycotoxin ochratoxin A (OTA) is produced mainly by mould species of the genera Aspergillus and Penicillium which can grow on cereals and other plants under a great variety of climate conditions. Cereal contamination with OTA has been most thoroughly studied, and reveals different levels of contamination around the world (JECFA 2001; EFSA 2006). OTA is also found in beer, coffee, cacao beans, peanuts, wine and in pork meat due to carry-over from feed (EC/SCOOP 2002b), and recent studies from Chile show its presence in cereal-based products and pork meat (Vega et al 2009; Muñoz et al. 2007).
OTA is a well known nephrotoxin in pigs, and was classified by IARC (1993) as possible human carcinogen (group 2b). OTA has been implicated in Balkan Endemic Nephropathy and associated with urinary tract tumours because of rather high OTA levels detected in food samples and in blood or urine from affected persons (Castegnaro et al. 2006). Exposure to OTA has been reduced in several countries over the years, by setting permissible maximum limits for OTA in foods and feeds (EC 2005; van Egmond et al. 2007). Nonetheless, humans continue to be exposed to this mycotoxin as a consequence of its widespread occurrence.
Exposure of humans to OTA through the diet and in occupational settings has been assessed by biological monitoring as reviewed by Scott (2005), and in more recent studies using blood plasma or serum and also urine samples (Aslam et al. 2005; Castegnaro et al. 2006; Degen et al. 2005, 2007; Pena et al. 2006; Zohair and Salem 2006; Vatinno et al. 2007). The toxicokinetics of OTA facilitate monitoring in biological fluids, in particular in blood. Once absorbed from the gastrointestinal tract, OTA binds strongly to plasma proteins which delays its renal clearance and results in a half-life of about 35 days (Studer-Rohr et al. 2000). Reabsorption of filtered OTA in the kidney, enterohepatic recirculation and slow biotransformation to more polar, excretable metabolites may further explain its long half-life in humans (Ringot et al. 2006). In the duodenum, OTA can be cleaved by intestinal microbial enzymes and mammalian proteolytic enzymes, e.g. carboxypeptidases (Pitout 1969; Doster and Sinnhuber 1972; Madhyastha et al. 1992), and converted to the non-toxic form, ochratoxin alpha (OTα). Recently, we identified OTα and its conjugates as major metabolites in human blood and in urine, along with unconjugated parent compound, OTA (Muñoz et al. 2009).
Analysis of Ochratoxin A in human milk samples from Europe
Concentration range (ng/L)
Gareis et al. 1988
Breitholtz-Emanuelsson et al. 1993b
Zimmerli and Dick 1995
Kovács et al. 1995
Micco et al. 1995
Miraglia et al. 1995
Skaug et al. 1998
Skaug et al. 2001
Turconi et al. 2004a
Postupolski et al. 2006
Galvano et al. 2008b
Dostal et al. 2008
Gürbay et al. 2009
The first study on OTA in human milk was that of Gareis et al. (1988) in Germany. They detected OTA in 11% (4 of 36) of the samples in a range of 17–30 ng/L. Low levels of OTA contamination were also reported in Switzerland (Zimmerli and Dick 1995). A higher frequency of mycotoxin-positive human milk samples was found in studies from Sweden and Norway which showed OTA levels in the range of 10–182 ng/L (Breitholtz-Emanuelsson et al. 1993a; Skaug et al. 1998, 2001). Much higher levels were detected in studies from the mid-1990s in Hungary (Kovács et al. 1995) or Italy (Micco et al. 1995; Miraglia et al. 1995), with maximal OTA concentrations of several micrograms per litre. In more recent studies by Turconi et al. (2004) and by Galvano et al. (2008), OTA concentrations in milk from Italian mothers were clearly lower, with 1–57 and 5–405 ng/L, respectively. Some of the studies also revealed a relationship between the maternal diet and OTA levels in milk: Skaug et al. (2001) found OTA more often in milk samples from individuals with a high intake of liver paste and cakes, and the risk of OTA contamination was also increased by the intake of juices. Galvano et al. (2008) reported that OTA levels were significantly higher in milk of habitual consumers of bread, bakery products and cured pork meat. Most recently, very high OTA levels have been reported for breast milk from Turkey (Gürbay et al. 2009).
Overall, it is apparent from these studies that OTA in human milk shows considerable individual and geographical variations. Considering the known toxicity of OTA, and the possibility of infants being particularly susceptible, it is clearly important to further assess their exposure (Degen 2000). In 2002, the European Community set a maximum level of 0.5 ppb (part per billion) in baby food (EC 2002a), and tolerable daily intake (TDI) values for OTA in adults have been proposed by different advisory committees (NNT 1991; JECFA 1996; SCF 1998; EFSA 2006). Studies in different countries show that, even at present levels of OTA in human milk, the TDI proposed for adults can be exceeded by infants due to their rather high food intake per body weight (e.g. Skaug et al. 2001; Galvano et al. 2008; Dostal et al. 2008). The Innocenti Declaration adopted by the WHO and UNICEF recognizes the need for protection, promotion and support of breast-feeding (WHO 1990). Consistent with this premise, the aim of this pilot study is to assess OTA exposure of newborns in Chile in the early stages of breastfeeding. Thus, we validated a method suitable for analysing OTA and its metabolite OTα in human milk and applied it to a first set of samples from Chilean mothers.
Materials and methods
Chemicals and materials
Methanol, chloroform, isopropanol, acetic acid (96%) and phosphoric acid (85%) were obtained from Merck (Darmstadt, Germany). The solvents used as mobile phase were HPLC grade. OTA (1 mg-standard, purity 98%) was purchased from Sigma (Taufkirchen, Germany) and OTα (11.9 μg/mL acetonitrile, purity 98.9%) from Biopure (Tulln, Austria). Strata C18 E (100 mg/1 mL) cartridges were used to perform the solid phase extraction (SPE). OTA was dissolved in methanol and calibrated spectrophotometrically at 333 nm using the molar extinction coefficient 6,400 M−1 cm−1. The working standard solutions were prepared weekly as dilutions in methanol/water (1:1, v/v) in a range from 0.10 to 10 ng/mL. For spiked samples, stock solutions of OTA and OTα (each 50 ng/mL) in methanol/water (1:1, v/v) were used. Enzyme ß-glucuronidase/arylsulfatase (β-Gluc/ArylS) from Helix pomatia (specific activity 5.5 U/mL ß-glucuronidase, 2.6 U/mL arylsulfatase at 37°C) was purchased from Roche (Mannheim, Germany), and used with 10-fold hydrolysis buffer (13.6 g sodium acetate hydrate, 1.0 g ascorbic acid, 0.1 g EDTA in 100 mL deionised water, adjusted to pH 5.0 with acetic acid 98%) for the enzymatic treatment of milk samples.
Collection of breast milk samples
For this pilot study, nine healthy women were asked to provide small amounts of breast milk (5–10 mL); the samples were collected at Higueras Hospital in Talcahuano City, Chile, during October 2008 and January 2009. The average age of the women was 30 ± 7 years (range 20–41 years), and their body weight ranged between 55 and 80 kg with a mean of 70 ± 9 kg. Milk collection was done within in a time up to 6 days after the delivery; one donor provided a second sample 28 days after delivery. The milk samples were stored at −20°C until analysis. The study was previously approved by the local ethic committee, and all mothers signed an informed consent before the collection of samples.
To 3 mL of human milk, 1 mL methanol/phosphoric acid (98:2, v/v) and 2 mL distilled water were added. The analytes (OTA and OTα) were then extracted with 3 mL chloroform. Some aliquots of milk were subjected to enzymatic hydrolysis with ß-glucuronidase/arylsulfatase (see below) prior to liquid–liquid extraction. Samples were centrifuged at 4000g for 10 min, and the top aqueous layer was aspirated with a pipette and discarded. Then, exactly 2 mL of the organic phase were transferred to a new vial and evaporated to dryness at 45°C under a gentle stream of nitrogen. The extract was reconstituted in 500 μL of methanol/water (1:1, v/v) and with 200 μL of n-hexane to remove the fat layer. The n-hexane layer (upper layer) was discarded and the samples were then filtered with a 0.45-μm pore size Teflon syringe filter. Milk extracts were analyzed by HPLC with fluorescence detection.
Enzymatic cleavage of conjugates
To account for the possible presence of phase II metabolites, i.e. glucuronide or sulfate conjugates, milk samples obtained in sufficient quantity were divided into aliquots. One aliquot (3 mL) was subjected to enzymatic treatment with β-Gluc/ArylS prior to further sample clean-up; the other aliquot (3 mL) was directly subjected to sample extraction (see above). An increase in the chromatographic peak of OTA and/or OTα in the enzyme-treated sample compared with the aglycone levels measured in the non-hydrolysed (parallel) sample is then indicative of the presence of conjugates. Initially, different amounts of ß-Gluc/ArylS enzyme solution were tested to ensure optimal conditions for cleavage of analyte conjugates. Suitable hydrolysis conditions were as follows: 3 mL milk were mixed with 0.40 mL hydrolysis buffer and 40 μL enzyme. All samples were kept at 37°C overnight prior to liquid–liquid extraction of total OTA and OTα.
HPLC analysis with fluorescence detection
The validation process was done using a breast milk sample from Germany with no detectable OTA contamination as blank. Cow milk from retail market was also tested using this method, but most of the cow milk samples analysed contained high natural OTα levels (confirmed by LC-MS) and a major interfering peak eluting close to OTα. In contrast, our cow milk samples contained only low levels of OTA. This is in line with findings by others who reported traces or non-detectable levels of OTA in cow’s milk (Boudra and Morgavi 2006; Boudra et al. 2007; Valenta and Goll 1996; Skaug 1999).
Recovery levels for OTα and OTA in human milk
Spike level (ng/L)
100 (n = 3)
500 (n = 3)
1,300 (n = 3)
Precision inter- and intra-day
Spike level (ng/L)
Inter-day (n = 7)
Intra-day (n = 8)
OTA and metabolite levels in human milk
Levels of OTA and OTα (aglycones) in milk samples from nine lactating mothers
A-a (Day 1)
A-b (Day 3)
I-a (Day 1)
I-b (Day 29)
Mean ± SD
106 ± 45
40 ± 30
OTA-transfer from maternal blood to milk has not been studied so far in great depth, and different factors, in addition to dietary intake, can probably affect OTA levels in breast milk, for instance the efficacy of OTA detoxication in the mother and the stage of breast milk sampling, i.e. early or later after delivery of the baby. Two mothers donated breast milk on two occasions: donor A gave milk at day one (A-a) and at day three (A-b) of breast feeding; donor I gave milk at day one (I-a), and 1 month later (I-b). The OTA levels for milk of both mothers at different sampling times are given in Table 4; the results for donor A indicate no major change in the OTA levels within a period of 2 days. OTA levels in milk of the second donor (I) showed a clear reduction from 184 to 44 ng/L, i.e. more than 70% within a period of 1 month. This may be related to the fact that right after delivery all components (including OTA) are more concentrated in colostrum than in mature human milk; on the other hand, we cannot exclude that the mothers dietary intake of OTA has changed in that period. Either way, the OTA concentrations detected in milk of Chilean mothers (Table 4) are within the range of values reported from Italy for human mature milk samples (Galvano et al. 2008) and for milk collected on the 4th or 6th day after delivery (Turconi et al. 2004).
Calculated OTA intake for newborns and infants with breast milk
A method developed and validated for analysis of OTA and its non-toxic metabolite OTα in human milk has been applied in a small pilot study in Chile. The method involves liquid–liquid extraction and HPLC analysis with fluorescence detection. All milk samples collected from nine mothers soon (up to 6 days) after delivery were found to contain OTA concentrations above the level of detection (LOD 10 ng/L), and average OTA contamination in milk samples was 106 ± 45 ng/L (range 44–184 ng/L). OTA concentrations were comparable with or without enzymatic cleavage of conjugates prior to extraction of milk, indicating that the mycotoxin is excreted as aglycone. The lack of detectable OTA glucuronides/sulphates in milk is in accord with recent data on their absence in human blood and urine (Muñoz et al. 2009). In contrast, the levels of its metabolite OTα in milk increased considerably after enzymatic treatment, indicating that the majority of OTα in milk is excreted as glucuronide or sulphate. The presence of OTα conjugates in milk was confirmed by LC-MS analysis (data not shown), and is in accordance with their presence in human blood and urine (Muñoz et al. 2009). Since concentrations of total OTα (208–843 ng/L) are higher than those measured for OTA (44–184 ng/L), it could be useful to include metabolite analysis in mycotoxin biomonitoring, despite the fact that OTα is of no toxicological importance in human milk.
These new data provide further insight into the fate of OTA in the human organism. On the one hand, they document partial detoxication of OTA to OTα and conjugation of this metabolite. On the other hand, the lack of OTA glucuronidation/sulphation is reassurance for the validity of OTA measurements in milk by other analytical methods, as no fraction of OTA has been missed that might release toxic aglycone upon cleavage in the organism. This also allows comparison of OTA levels in milk reported in other countries (Table 1) with values found in human milk from Chile (Table 4). By and large, OTA levels in milk of Chilean mothers were within the range of values reported from Norway (Skaug et al. 2001) or from Italy for human mature milk samples (Galvano et al. 2008), and for milk collected on the 4th or 6th day after delivery (Turconi et al. 2004). A comparison of average OTA concentrations in breast milk samples from Chile to blood concentrations reported in women from two regions (0.44 and 0.88 ng/mL serum; Muñoz et al. 2006), is indicative of an efficient lactational transfer of the mycotoxin.
In light of the known toxicity of OTA, it is of interest to compare now infant exposure with milk to tolerable daily intake (TDI) values derived in risk assessments for the mycotoxin. The Joint FAO/WHO Expert Committee on Food Additives (JECFA 1996) proposed a provisional tolerable intake of 100 ng/kg bw/week, equivalent to 14 ng/kg bw/day, and the European Food Safety Agency (EFSA 2006) has recently established similar values, up to 120 ng/kg bw/week, equivalent to 17 ng/kg bw/day for adults. Both JECFA and EFSA chose nephrotoxicity in pigs as the critical endpoint, and applied a set of uncertainty (safety) factors to account for interspecies differences in toxicokinetics and toxicodynamics and sensitive subpopulations. Other expert groups, e.g. the Nordic Working group on Food Toxicology and the Scientific Committee for Food, have previously proposed lower TDI values of 5 ng/kg bw/day (NNT 1991; SCF 1998), and Canadian authorities suggested a permissible daily intake of 1.2–5.7 ng/kg bw (Kuiper-Goodman 1991). They chose carcinogenicity in rodents as critical endpoint and applied higher safety factors to account for uncertainties in the database (Degen 2000).
Choosing an appropriate TDI value for newborns/infants is not trivial for several reasons: (1) in infants detoxifying enzymes are less developed, and glomerular filtration is about 1/3 lower than in adults (Postupolski et al. 2006); and (2) aside from OTA, breast milk is also a potential source of other nephrotoxic agents, for instance cadmium and lead (Turconi et al. 2004) or citrinin, a known co-contaminant of OTA (Krogh et al. 1973; Vrabcheva et al. 2000). This may justify comparing the calculated daily intake for infants in Chile (Fig. 3) with the TDI of 5 ng/kg bw, rather than a TDI of 14–16 ng/kg bw set for adults. With a few exceptions (milk of donor A-b and I-b), OTA intake of infants more or less clearly exceeded the proposed TDI value of 5 ng/kg body weight at early stages (day 6) of breastfeeding. On the other hand, comparing OTA-intake values for sampling times 28 days apart revealed an exposure with the first milk sample (I-a) of 15 ng/kg bw/day (about 3 times the TDI value), whilst OTA intake with milk of the second sampling (I-b) dropped below the TDI. This example points to a significant decrease in OTA-intake levels within a period of 1 month. When choosing the higher TDI of JECFA and EFSA, infant exposure still approached the provisional tolerable doses of 14–16 ng/kg bw/day proposed by JECFA and by EFSA for adults.
In conclusion, first data on OTA in human milk in a small cohort in Chile show the presence of this mycotoxin at levels which may pose a risk to the developing infant. This warrants further investigations and follow-up studies on the levels of OTA in milk at later stages of breastfeeding. Such studies should also include parallel analysis of maternal blood and milk to investigate the lactational transfer of OTA and possible changes over time. With regard to risk management, it is worth considering dietary recommendations for nursing mothers. Moreover, setting limit values for OTA in food and monitoring in countries with no existing mycotoxin regulations is recommended.
The authors want to thank the midwives Verónica Moreno V. and Barbara Rubilar R. from the Higueras Hospital in Talcahuano, Chile, and Cecilia Urrutia for their help by collecting samples, and Iris Glaeser, Michael Porta and Gabi Baumhoer (IfADo) for their technical help. Finally, we want to thank DAAD and CONICYT for their support by a stipend to K.M.